Abstract
Observational studies of the Pacific basin since the 1950's have
demonstrated that a decrease (increase) in tropical Pacific sea
surface temperatures (SSTs) is significantly correlated with a spin-up
(slow-down) of the Pacific Subtropical Cells (STCs). STCs are shallow
wind-driven overturning circulations that provide a pathway by which
extratropical atmospheric variability can impact the equatorial
Pacific thermocline and, through upwelling in the eastern equatorial
Pacific, tropical Pacific SSTs. Recent studies have shown that this
observed relationship between SSTs and STCs is absent in coupled
climate model simulations of the late 19th-20th centuries. In this
paper we investigate what causes this relationship to breakdown and to
what extent this limits the models' ability to simulate observed
climate change in the equatorial Pacific since the late 19th
century. To provide insight into these questions we first show that
the NCAR Community Climate System Model's simulation of observed
climate change since the 1970's has a robust signal in the equatorial
Pacific that bears a close resemblance to observations. Strikingly
absent is a robust signal in the equatorial thermocline due to the
underestimate of STC variability. Our results suggest that the coupled
model may be reproducing the observed local ocean response to changes
in forcing but inadequately reproducing the remote STC-forcing of the
tropical Pacific due to the underestimate of extratropical winds that
force these ocean circulations. These conclusions are found to be
valid in 5 different coupled climate model simulations of the late
19th-20th centuries (CCSM3, GISS EH, GFDL-CM2.1, CSIRO-Mk3, and
HadCM3).